Research map / 3D vision / Depth and stereo
Towards Interpretable Deep Networks for Monocular Depth Estimation
Zunzhi You, Yi-Hsuan Tsai, Wei-Chen Chiu, Guanbin Li
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| assign a depth range to each unit | ResNet-50 MDE network · NYUD-V2 | depth selectivity (testing): 0.4286 -> 0.7529 | Improved |
| assign a depth range to each unit | SENet-154 MDE network · NYUD-V2 | depth selectivity (testing): 0.4691 -> 0.7693 | Improved |
| assign a depth range to each unit | ResNet-50 MDE network · NYUD-V2 | RMS: 0.443 -> 0.422 | Improved |
| assign a depth range to each unit | SENet-154 MDE network · NYUD-V2 | delta<1.25: 0.874 -> 0.882 | Improved |
| assign depth ranges instead of directly regularizing selectivity (ablation) | ResNet-50 MDE network · NYUD-V2 | depth selectivity (testing): 0.6039 -> 0.7529 | Improved |
| assign depth ranges instead of directly regularizing selectivity (ablation) | SENet-154 MDE network · NYUD-V2 | depth selectivity (testing): 0.5694 -> 0.7693 | Improved |
| assign depth ranges instead of directly regularizing selectivity (ablation) | ResNet-50 MDE network · NYUD-V2 | delta<1.25: 0.857 -> 0.861 | No clear effect |
| correct unit responses using ground-truth depth averages | baseline ResNet-50 · NYUD-V2 | delta<1.25: 0.849 -> 0.779 | Hurt |
| correct unit responses using ground-truth depth averages | interpretable ResNet-50 · NYUD-V2 | delta<1.25: 0.861 -> 0.947 | Improved |
| apply method to CSPN depth completion | CSPN · NYUD-V2 depth completion | depth selectivity (testing): 0.4213 -> 0.9475 | Improved |
| apply method to CSPN depth completion | CSPN · NYUD-V2 depth completion | RMS: 0.117 -> 0.119 | Hurt |
| apply method at layer upconv1 of DenseNet-161 model | DenseNet-161 MDE · KITTI | delta<1.25: 0.963 -> 0.959 | No clear effect |
Closest papers by meaning
- Network Dissection: Quantifying Interpretability of Deep Visual Representations (CVPR 2017 · similarity 0.88)
- Visualization of Convolutional Neural Networks for Monocular Depth Estimation (ICCV 2019 · similarity 0.88)
- Interpreting Deep Visual Representations via Network Dissection (TPAMI 2018 · similarity 0.87)
- Look Deeper into Depth: Monocular Depth Estimation with Semantic Booster and Attention-Driven Loss (ECCV 2018 · similarity 0.87)
- Measuring Per-Unit Interpretability at Scale Without Humans (NeurIPS 2024 · similarity 0.86)
- DepthCues: Evaluating Monocular Depth Perception in Large Vision Models (CVPR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
